Multitasking—defined as the more or less overlapping execution of two or more tasks—has been shown to impair performance and induce performance costs. Nevertheless, individuals frequently engage in multitasking for private and work-related purposes. This raises the question: Do subjective impressions about multitasking efficiency diverge from objective measures, or can multitasking indeed comprise benefits in certain conditions? While previous assessments of multitasking efficiency focused mostly on central processing limitations, they largely neglected benefits of parallel peripheral task processing in overlapping task execution. To address this empirically, we conducted two multitasking studies and compared Time on Task (ToT; Reissland & Manzey, 2016), that is, the total time needed to finish both sub-tasks, in an overlapping versus a sequential dual-task condition. Experiment 1 (N = 18) applied a Psychological Refractory Period (PRP) paradigm with distinct task sets in Task 1 and Task 2. Experiment 2 (N = 32) used identical task sets to maximize between-task interference in the overlapping dual-task condition. Results of both experiments showed shorter ToT in the overlapping compared to the sequential dual-task condition, without much evidence of compensatory effects. These results suggest that the assessment of multitasking efficiency should consider the total time on task needed to perform both tasks to capture the costs of central processing limitations as well as benefits of parallel peripheral task processing. They further underscore the importance of carefully selecting and justifying the chosen evaluation metrics when assessing multitasking efficiency.